Speaker 1: 0:23

Welcome to Techzine TV.

Speaker 2: 0:27

Welcome to Techzine TV. We're at Workday Rising in San Francisco. We're talking to Peter Deiles. He's the CTO of Workday. Welcome, peter, talking about strategy, because that's basically, if you're part of the management team, that's one of your big roles. You're also a pretty technical guy. I think you want to manage money and people, which you already did, but you also want to manage AI agents with the agent system of record, something you announced during the keynotes. Yeah, can you explain a bit what you mean by agent system of record? Yeah, it's a great question.

Speaker 3: 1:05

So Workday is the system of record for people, for our customers. So, basically, who's employed, what department do they work in, who's their manager, what's their entire employment history, job change, everything there is there, and it's not just for HR. Actually, when you log into your identity provider let's say your customer Workday and Okta, okta is using Workday data to go figure out you know, are you employed, where do you sit in the org chart, so on. Right, so Workday is actually like the database behind Workday is the system for people. Now, when you think about AI agents, I think the real trick of agents and the real breakthrough is people figure out. How do you go and incorporate AI in your company.

Speaker 3: 1:49

Well, actually personifying AI through an agent, like a digital coworker that you can today chat with, over time will be doing more tasks on your behalf, like that interface, is actually incredibly convenient. How far do you want to take that? I mean, I think we'll see where the foundation models go and, from the agent-system of record perspective you're going to personify, have these entities that are working on your behalf the same construct you have for people for the people system of record? Extending that to agents is actually quite nice. So, what department is this agent a part of. I can assign it an ID through Workday or through something like Entra, which you now saw on stage with Microsoft, and then, from an IDP and identity perspective getting a little nerdy here the protocols for authenticating people already exist for that, so being able to have a system of record for agents is actually really convenient in terms of evolving how AI and people work together, with an existing set of protocols, an existing database, an existing way of.

Speaker 2: 2:51

So those are the goals and capabilities. Yeah, exactly, but you can do that on your own platform. But there are going to be massive amounts of platforms with. Ai agents you want to control them externally as well.

Speaker 3: 3:02

Yeah, it's great. I mean, look, there's a massive number of SaaS applications as well. Right, no-transcript about saying, if you want to register and have these in one place with a way that you already managed one of your most sensitive assets in your company, which is people you know, this is a pretty convenient way to do it and so far you, the ecosystem response is pretty good because people realize they're not going to be the only game in town, whether it's agent force or co-pilot or I'm using, um, uh, agent space. We had, we had Google on stage as well. Uh, richard from Google.

Speaker 3: 4:00

So like it's kind of this realization that it's going to be a not-size-fits-all, not a winner-take-all, but having a single place to go register these and manage them is pretty compelling. It helps our customers who are trying to make sense of all of the. You know, how do I go govern all of this? How do I enable open adoption of these things but also have a little bit of semblance of control and security? Asor is a very useful contract for that of control and security. Asor is a very useful construct for that.

Speaker 2: 4:27

But do you think there will be a standard or a protocol for managing agents across platforms?

Speaker 3: 4:33

Absolutely. Yeah, I think there will be. Today it's very early. So MCP, you know widely adopted Dirty secret a lot of agents, actually, you know, instead of using something like the agent-to-agent protocol, they're just seeking MCP as well, and so juries out on how that evolves.

Speaker 3: 4:49

What A2A has from a protocol perspective, which is incredibly valuable, is a description of the agent, which an MCP doesn't give you. That at all. You don't have a here's my skill. You don't need a protocol for a hypervisor level to manage it. Exactly right, and so I think this will evolve and extend, and it's something where you know we're in the ecosystem, we're A2A partners, one of the areas that we're really excited about, which is very simple but incredibly impactful Agents often work on behalf of a user, so being able to have delegate permissions for agents is really important.

Speaker 3: 5:22

You can do that in the. You know the OAuth standard, right, but you've got to get agreement from everyone who's accepting these tokens that, hey, you know, this agent's on behalf of Peter. That's the type of tweak where I think you'll actually see incremental evolution, mcp being maybe the first starting point, and then you know we're along for the ride. We really do believe you know our right to win and where Workday's best is in managing the people and money and future agents and we're gonna work with everyone to go and build the best protocols wherever people wanna create agents.

Speaker 2: 5:53

Okay, but you're also buying those companies at the moment. Well, and I noticed, you now have three local platforms in your portfolio at least, or in your technology stacks. So you're a lot about building agents and you're buying these companies for a reason for talent, but also for technology. You want customers to build apps and agents on top of your Workday platform, because you call it a platform now instead of applications. How far is that that that's ambition or objective? Yeah, to be the platform for everything it's a great question.

Speaker 3: 6:34

I mean, let's go back to what we've bought recent acquisition. Right, we require flow wise, which is an open source toolkit. It's an open source platform. Uh, 40 000 stars in gith GitHub for building drag and drop agents.

Speaker 2: 6:48

For people who don't know it. It's kind of a competitor for NHN. It's a bit more popular, sorry.

Speaker 3: 6:56

They're up there. Nhn, I think, is also playing more in the Zapier style space as well, doing more data integration as well. It's like many flowers bloom. Perhaps we felt like, hey, our customers want to build more agents on Workday. We can go and build our own UI. We can go and integrate with N8n, which we will continue to do. We can go and integrate with Flowize. But also, you know, we think the Flowize team is great. We want to pour fuel on the Flowize team and grow. So we're just going to keep FlowWise open source. Our customers will use FlowWise in Workday because it can be. You know best, on Workday you want to build with people and money data, so we're not building, you know, generic. You know Microsoft co-pilot agent space, so on. You know we're like our customers spend so much time in Workday. This is the most sensitive data to go feed their AI. They can do that in our security perimeter. That's great.

Speaker 3: 7:45

On the other two acquisitions we made for Paradox, this is about recruiting. They built an incredible recruiting section. I mean they do 31 million sort of candidate screens a year. That's one per second. It's an incredible scale, huge overlap with the Workday customer base. They have their own ATS, but they also work with Workday. So it's just that's like, hey, we can go build our own recruiting agent suites and we have some cool stuff going there. But like we can just, you know, we have deep partnership with the company. We think there's a better together story.

Speaker 3: 8:13

Our customers love buying just one platform. It's so painful to go buy many different, so we'll let these partners proliferate. But we want to make it easier. So that's for recruiting. And then on the sauna side, this one is about kind of two things. One, they have a great learning product, yeah, but two, the bigger story is they have an amazing ui for ai. So as we've thought about how does workday experience evolve? You know we were very close to the sauna team already. We said this is a really incredible opportunity to take a lot of the work already doing in workday on our agent and surface this through a really beautiful consumer grade UI. So no code platform yeah, somewhat. I think about Sana as essentially connecting all it's enterprise search.

Speaker 3: 8:53

It's like all of my different systems, which is, in some ways, no code, but really just about Everybody is into enterprise search. Well, I think it's an early market. To be honest, I think that, you know, when you look at the actual penetration and use of this, it's still very, very early. Right, I think it will be the future way people work with their enterprise data, but now we have an amazing enterprise search platform. Of course, on top of enterprise search, what do you want to do? You want to stitch together multiple different systems. You can do that by statically coding these or by doing a UI, or you can have an agent going, so it's really different.

Speaker 3: 9:28

Personas Like FlowWise is our low-code agent builder. It's open source, extend is our pro-code build-your-own-app-insider workday, and Sana will be the experience for less technical users to go and use their enterprise data. So it all kind of fits together, I think. What is fair, though, is we are on a tear. You know like in the last six months, the team's shipping of first-party agents has just gone through the roof. We announced a massive number across the suite.

Speaker 2: 9:55

Yeah, you just took all your expertise and built agents for HR and for finance functions that people have a hard time or requires a lot of manual work. That's automated with agents.

Speaker 3: 10:07

Yeah, and a ton of delivery. And then now on the acquisition side, these are one recruiting agent, two other platforms to go and accelerate that vision. So I think you know, I don't know that we'll have the same, I don't expect we'll have that same. You know number of acquisitions at Next rising in a year. But I think that we are laying the foundation for radically different work.

Speaker 2: 10:27

Yeah, what I wanted to say was you're building or you're buying companies that help you give your users the tools to build applications off your platform. Yeah, that's kind of the technology that you're buying and integrating. Yeah To have more, yeah To have more. And I also spoke to Carl yesterday and he also said like yeah, we want to be the primary platform for organizations to build a region.

Speaker 3: 10:55

Well, yeah, I think that's right. You know our right to win here is people and money, and you know we have 75 million users, so that's like an amazing opportunity to go and distribute. We think at the future of work.

Speaker 2: 11:12

But if you look at the market, everybody's doing enterprise search now. They're all launching their AI platforms where they can build agents. Are you really going to try to compete with the service now and the sales forces and the SAPs as well? I think we have to.

Speaker 3: 11:30

I think we will, isn't that a?

Speaker 2: 11:32

bit too big a mission.

Speaker 3: 11:34

You know what we have our people and money. We've got the best data set out in the world. You know my experience building AI for one of the largest hyperscalers in the world is actually that you know the models can get better and better. We've at GDM making, for example, codegen for data great. The thing that actually makes AI work super, super, super well is context. So if I can understand, you know, what is this user trying to do? What are users like them trying to do? What is going on in their organization? I can provide that context, especially the latest models, and get way better results.

Speaker 3: 12:09

And so the bet, which I'm pretty confident in I put money where my mouth is yesterday is that we can go and build really good enterprise search and, frankly, I actually think that the distribution advantage that we have with our customer base, who is so hungry for AI they are so excited to see what does the future look like and then to give that to them via the platform that we announced yesterday. You can just turn this on and you pay for what you use, but there's no Workday Pro, there's no extra platform fee, there's nothing. Just to say, look, try it and then, if incumbent. What I'm excited about. This is what I get so fired up about. We get to go and ship and go build that and the team Asana team these guys are really good. They're really good at the ML side and they're also extremely good at the user experience. Foundation models get better. If you own the experience layer, I think you win the warf.

Speaker 2: 13:08

I think you're onto something, but I do think the ambition is too high and I'll tell you why. Because I do think it are great acquisitions and great products. The Sun looks great. I didn't know it before, but I did look it up, of course, and it looks really nice and your strategy aligns and there's definitely something there.

Speaker 2: 13:27

But if we take a step back from an enterprise organization perspective, they are going to choose, I think, a primary platform to build their AI agents. They all buy Salesforce, they all buy Workday, they all buy ServiceNow, they all buy SAP. Most enterprises have all. Basically. So I think they're gonna base their choice to build AI agents on where their business lies. So if they are manufacturer using a lot of ERP SAP, for example they're gonna choose them. If it's a sales or service organization, they're more likely to go to Salesforce, and if it's I don't know they've been building workflows and service now for the last 10 years it's easy to take that platform. If we look at Workday, workday is not in their primary business. It's a supporting role on HR and finance. You do need HR and you do need finance, but it's more supporting. It's not the primary group. That's why I think they will more likely choose the other ones over urban.

Speaker 3: 14:34

Yeah, maybe. What do you think? It's an open question when you think about these agent platforms. Let's talk about generic agents for work. How do I go and stitch together my enterprise data source? How do I build workflows? How do I build configuration about how they can go and share, like the foundation model, companies are ultimately going to dictate how powerful those are. You know, very few people outside of maybe doing retrieval are going to fine-tune models for this task. All about foundation model enterprise connectors.

Speaker 3: 15:04

The enterprise connector perspective is people, about more mcp. You have better federated search. You know the advantage essentially becomes much more. How do you build the best experience layer on top of foundation models Plus essentially federated search with the universal ACLs and that's evolving super fast. And so, from that perspective, if I'm in manufacturing, you know, am I going to go use one of our competitors, my primary system for this? Potentially, if I'm going to go use my sales stuff, my know-how at Salesforce, I should let many flowers bloom. But here's the thing If we have the best experience layer, every single employee for every single working customer logs into Workday, they take time off, they check their paychecks, they get promoted, they write performance reviews.

Speaker 3: 15:48

So I think that that advantage where you're wall-to-wall in a company. I'm not just looking at it from a FP&A perspective, I'm not looking at it from a manufacturing perspective. We're not like a narrow SaaS application. We are wall-to-wall inside of an organization where you have HCM, Even for FINS, you have the full system of record, the general ledger for the company. I think that's a unique advantage because it's basically the company operations, the most important part of the company, the employees of the company and the money for the company that's in Workday. So that's the wedge. You know it's across industries. We have 11,000 customers, unbelievable customer base distribution.

Speaker 3: 16:28

Well, I actually joined Workday four months ago because of the customer base and that tells me if we can build the best product, the best product experience and we can distribute it. I think it's an incredible chapter for Workday. Okay, it's not traditional Workday. I think you know maybe the question. The question is what is Workday doing? Why is Workday buying Sana? What is? How does Workday think? I get it? And and and, but it's a new. It's a new thing.

Speaker 2: 16:58

I even yeah, because you're an AI guy. I asked Claude to do research on your financials. I say, take that line of employees, the financials, can they even compete with these guys? And Klaus came up all by itself. It's a David versus Goliath battle. You're five times smaller than the average other platforms.

Speaker 3: 17:22

Maybe I think it comes down to delivery. That's true. If an employee account was everything, you would have no startups, no innovation. That's true. Like if employee count was everything you would have, you know, no startups, no innovation. That's true.

Speaker 2: 17:34

But efficiency you now have. You're a very efficient organization. If you look at the financials, for sure. If you look at revenue per employee, yeah, if you're going to expand that big, it's really hard to keep it.

Speaker 3: 17:45

No, but I think you're missing the point, right? Yeah, like the future of software is built by cracked teams of software engineers running on the latest coding. You know all, basically cloud code and codecs, and then you, you know, talk to Joel. You know how he's thinking about expanding his org. I don't want more people. They are hiring more. They have to hire more, but it's a bigger team. They're not better teams, especially with AI.

Speaker 2: 18:09

I mean it's a couple of developers right last year that was somewhere in the keynote.

Speaker 3: 18:15

No, no, not our developers. So that is the number of developers on our platform. Yeah, we double-numbered developers on the platform by opening up and actually I think that's the real message is that you know, workday, we can try to build it all, especially if you think our aspirations in Medium Enterprise or aspirations in International, our aspirations to go and do more for knowledge workers, the more we can express open standards so you can just code and cloud code, codex, all the agents know how to go and use Workday. That's fantastic. It's fantastic for our customers, fantastic for Workday as a business, fantastic for our partners. So the doubling in developers building on Workday. But I can tell you our dev teams have not doubled, they've gotten way faster.

Speaker 3: 18:54

You know, week two, my employment at Workday we rolled out Cursor for everyone. We had like 2,000 people in the company coding on Cursor in, like you know, one month from our pilot. So like Workday is shipping like ever before, and for Sana in particular, you know they've got a big, big task ahead of them. You use the product and you can just sign up and use it. Today it's really good, like it's a beautiful experience. I think maybe this is like something in the water of Sweden, or maybe by virtue of being out in you know, not in Silicon Valley, but these guys really, really care about their craft. It shows through the product experience. I think, from an end user perspective, it's the best enterprise search, best knowledge management platform on the market today. That's why we bought them for a billion dollars.

Speaker 2: 19:42

Yeah, we're going to see how it all turns out.

Speaker 3: 19:47

And the thing is this you talk about these competitors, you talk about them. You know a couple of names, you know. I think that the ambition we have as a management team at Workday is massive. Right, we're going to run the table on people and money and it all comes down to execution.

Speaker 3: 19:59

So this, this rising was here's where we're going. Here's this strategy. It's a older than we've ever been before. We sat with the analysts earlier today product games covered. For years. We've never seen the confidence here and the reality is, you know, we've got a lot cooking on our own product teams. You know we saw the agents that are coming out and the way we're going to commercialize it and basically give it away. Try it for free with your flex credits. And then we got the ambition, these acquisitions I think you know it's recognized as some great talent where they can accelerate our vision by 12, 18, 24 months and, by the way, bring that founder mode energy to this company.

Speaker 2: 20:38

Yeah, I would have expected you to say let's embrace the supporting role and integrate with everybody else.

Speaker 3: 20:44

But it's both. So, look, we're going to build the best agent for people tonight. That's our right to win. Okay, okay, 100%, 100%. And on top of that, if you don't like our agents for people and money or you want to use that data, here's your open data. Here's the Workday data cloud. It's basically Iceberg. You know Workday's never had a, you know, an open platform like that.

Speaker 2: 21:05

No, I know because I spoke to some of your customers and they were pretty annoyed by the way, integrating with Workday and get the data I mean you can get the data out.

Speaker 3: 21:13

We've always been pretty hard, but doing this in a way with open standards. And I think the key thing with iceberg is secure. So it's zero copy running in our cloud, you know, and we can apply the same security policies. So that was, I think, the rub was that without the technology really to do super secure querying over this, you know it was just copy everything out. So it's better for customers on two fronts One, it's zero copy. Two, you know it scales right. This thing is like an awesome. We had Sridhar from Snowflake on stage about this right?

Speaker 2: 21:43

How do we treat partnerships with a zero copy? Are there more to come?

Speaker 3: 21:46

You know we'll see, but I think there's a lot of excitement from the ecosystem because here's the thing, like. Because here's the thing like. We have so many CIOs attending this event this week, and the reason why is not because they're, you know, suddenly HCM zealots okay, even present company included. I was a Workday user but I didn't know anything about HCM. The reason why the CIOs are here, the reason why I'm here, is that when you think about AI, you need data to go and feed these models, and Workday is where this organizational data lives. So we're going to build kick-ass agents and kick-ass experience for those agents on our platform. But if you want to build your own or you want to go integrate, salesforce is one of our launch partners.

Speaker 3: 22:22

You mentioned Salesforce. They're partnering on the Workday data lake, because guess what? Salesforce data plus Workday data together way stronger. And you know you can go from Salesforce to Workday, you can go from Workday to Salesforce, you can go from Databricks and query both. But we you know it's like this is the next generation of SaaS Like we will always be that source of record for people and money. We have this enduring moat here and for our customers, they trust us with this data as they should, but we're making it easier to go use all of that across the stack and we're going to play with everyone in the space, including other knowledge management systems.

Speaker 3: 22:56

You know, if you want to do enterprise search on Workday, knock yourself out Like we're going to have the best enterprise search experience. That's my bet with Sana. But if you want to go around another one, you've got the APIs to do it and that's what's so fun about this era. Either way, our customers win. They have the choice. And on the product side, internally on the first party agent like the team is just incredible because we have this deep expertise on people and money and we offer some incredible talent on the AI side.

Speaker 3: 23:24

And you put two in a box and say you know what can you go and build? How do you reimagine payroll? How do you do a performance review with an agent? How do you go? By the way, it's an enterprise search problem. Turns out, performance reviews are an enterprise search. Instead of saying, oh, what did I think about John the last week, I can go pull on everything I know about John Enterprise search. So this is all converging. And one final piece the protocols here to power all of this. That's where I think the action is going to be. How do you make this stuff easier, more turnkey? Today, mcp is so rudimentary, it's like not even TCP, it's like IP. You know, we're like so low in the networking stack. I can't wait for the HTTP for enterprise SaaS and the AI. That's what we got to work towards.

Speaker 2: 24:01

HTTP for enterprise SaaS. Okay, you only have four months. Yeah, how much of this strategy comes from you and how much is already there?

Speaker 3: 24:11

Look, it's a team effort.

Speaker 2: 24:13

There's a lot of great ideas, but I'm getting the feeling some of this is your thing, because I think they got you for a reason.

Speaker 3: 24:21

You know, it's been really great learning from the workmates, kind of what's worked well, what hasn't worked well. I think there's a lot of ideas people have had. But, you know, going out and being as bold as saying we're going to have an open data lake, yeah, because the open report.

Speaker 2: 24:35

they're really close, so someone had.

Speaker 3: 24:38

Yeah, you know, I think we, you know I give the workmates the credit for it, but I'm happy to use the authority I have as CTO of Workday to say, yes, let's go and do this, let's go full throttle on this, right, and it's actually. You know, that's the air cover that we're providing the teams who are going and building and building this out, and I just spend a lot of time looking at the architecture, make sure it looks great, but there's some really, really talented, you know, work that maybe historically is a little more closed than we've talked about. It's always been open in integration sense, but not the same Embracement of Open Standard. You know the work made to get it. I think it was more. You know how do we?

Speaker 2: 25:15

I'd probably even call it a wallet garden.

Speaker 3: 25:17

so I won't use that phrase but I will say, if we were, you know we're knocking down the walls and that I think you know I won't take credit for it, but I'm happy part of the team that's doing it.

Speaker 2: 25:30

Okay, thank you very much for this conversation.

Speaker 3: 25:33

Yeah, I'm excited to see you next rising when we see this next wave of enterprise search play out. I think your questions are spot on and you know the reality is all about execution in this space and when I look at the landscape and how our fellow players in the ecosystem are evolving, the bet we're placing on Warp Day is we can ship faster, we can deliver to customers better product, and the proof is in the pudding.

Speaker 2: 26:00

You talked about interface. You said SANA is the best interface. Microsoft is betting on Teams, not a great interface. Salesforce is betting on Slack Maybe also. Not for everything. That's the interface. Salesforce is betting on Slack, Maybe also not for everything. That's the interface, but it are applications where people live it through the day. People are not so much living your workday, so you have to change that.

Speaker 3: 26:27

Not at all. Yeah, no, I mean look the Asana customers using the agent product. Their new tab is Asana, and the question I pose to you is what's the new browser tab for most knowledge workers?

Speaker 2: 26:41

It's not Slack, it's not Teams.

Speaker 3: 26:43

I think the browser industry is also going to change Massively, massively, but there will be some front door and a new front door. Chat is kind of a modality that sort of works today. It's probably not the end game modality by any stretch and I think that like communication channels, look, we have workday everywhere. We've got, we should, we had microsoft on the stage, we showed you know workday showing up in the code. It really comes back to this like we're going to compete on the enterprise experience layer because it's it's, that's blue sky. No one's doing it because you know it's kind of funny. You ever talk to vcs like a couple years back. You know enterprise searches, it's like a nerd candy, like such a fun problem to solve. Yeah, google even sold an enterprise search appliance. They put a round. We tested one we.

Speaker 2: 27:27

We bought one from ebay, yeah yeah to see how it worked, because it was. It was really expensive. It's incredible, but we wanted to know as journalists how does it work and what does it do? Because you would have expected, with all the technology Google had and else, that they would have covered enterprise search a year ago.

Speaker 3: 27:46

And I think pre-LLMs just lack of training. Data ambiguity in queries you need LLMs to be able to go and bridge that on a smaller data like customer regime. So now suddenly this category was like doa. You know you go to any venture capitalist, you know these guys fund craziest stuff. They fund fusion reactors, right, they wouldn't touch enterprise search. It was always like thank you, that's such a cool idea.

Speaker 3: 28:09

You know it's a great idea why don't you look at, like you know, financial planning instead? Okay, so, so like now that's real and it's here and there's products that are incredibly widely adopted, but regardless of that, like the actual average knowledge worker or even the early adopter knowledge worker, like that is still, you know, once they're front door for work, it's still blue sky. So the tech is now viable, but the market is wide open and I will be shocked if the front door for work is, let's say, group communication platforms as we know them today. We're betting it'll be Sana and as Sana evolves in the future, it's going to be Of course, yeah, and Sana is collaborative At Sana.

Speaker 3: 29:02

You can invite other people, you can share workflows, you can create content. I mean Sana runs everything from their all-team meetings to their stand-ups in Sana. It's a multiplayer. I mean, that's kind of what I love about it. Figma Figma shouldn't have existed, but they built the best experience on top of basically like WebGL and the Camas layout. The bad is that the Sana team is that and then you couple that a merging company with basically the distribution of all of the Workday customers and Workday users. It's going to be a crazy chapter and really fantastic for customers, but we got to keep delivering and that's what you should expect more of. Okay.

Speaker 2: 29:45

Let's have a conversation again in the future. I see, sign me up, I'll follow you, we'll see. Yeah, thank you all for watching TechSite TV.

Speaker 3: 29:53

Thank you so much.

Speaker 2: 29:54

And keep watching. All right, thanks, thank you.

Speaker 1: 30:01

Thank you for watching TechSign TV, the channel about enterprise technology that brings you IT insights and analyses from events all around the globe. We cover everything everywhere. Visit TechSigneu for more written in-depth articles and analysis, or keep watching techscenetv. Don't forget to subscribe to our channels and share your favorite videos with your colleagues. We'll see you soon.